ragora
Search your knowledge bases from any AI assistant. Ragora is a knowledge marketplace that lets you upload, organize, and search documents using hybrid RAG (dense + sparse vectors). Connect via MCP to search across all your collections directly from Claude, Cursor, VS Code, and more.
README
Ragora MCP Server
Search your knowledge bases from any AI assistant using the Model Context Protocol.
Ragora's MCP server is a hosted cloud service — no local installation required. Just configure your AI tool with the server URL and your API key.
Server URL: https://mcp.ragora.app/mcp
<a href="https://glama.ai/mcp/servers/@velarynai/ragora"> <img width="380" height="200" src="https://glama.ai/mcp/servers/@velarynai/ragora/badge" alt="ragora MCP server" /> </a>
Quick Start
1. Get an API Key
Sign up at ragora.app and create an API key at Settings > API Keys.
2. Connect Your AI Tool
Claude Desktop
Option 1: CLI (recommended)
claude mcp add --transport http ragora https://mcp.ragora.app/mcp \
--header "Authorization: Bearer sk_live_your_key_here"
Option 2: Config file
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"ragora": {
"type": "http",
"url": "https://mcp.ragora.app/mcp",
"headers": {
"Authorization": "Bearer sk_live_your_key_here"
}
}
}
}
Claude Code
claude mcp add --transport http ragora https://mcp.ragora.app/mcp \
--header "Authorization: Bearer sk_live_your_key_here"
Cursor
Add to .cursor/mcp.json in your project root:
{
"mcpServers": {
"ragora": {
"type": "http",
"url": "https://mcp.ragora.app/mcp",
"headers": {
"Authorization": "Bearer sk_live_your_key_here"
}
}
}
}
VS Code / GitHub Copilot
Add to .vscode/mcp.json in your project root:
{
"mcpServers": {
"ragora": {
"type": "http",
"url": "https://mcp.ragora.app/mcp",
"headers": {
"Authorization": "Bearer sk_live_your_key_here"
}
}
}
}
Windsurf
Add to ~/.codeium/windsurf/mcp_config.json:
{
"mcpServers": {
"ragora": {
"serverUrl": "https://mcp.ragora.app/mcp",
"headers": {
"Authorization": "Bearer sk_live_your_key_here"
}
}
}
}
Any MCP Client
Ragora uses Streamable HTTP transport. Point any MCP-compatible client to:
- URL:
https://mcp.ragora.app/mcp - Method:
POST - Header:
Authorization: Bearer sk_live_your_key_here
Available Tools
| Tool | Description |
|---|---|
discover_collections |
List all knowledge bases you have access to, with descriptions, stats, and usage examples |
search |
Search across all your knowledge bases at once |
search_collection |
Search a specific collection by name, with optional tag/metadata filters |
Tool Parameters
search(query, top_k?)
query(required): Natural language search querytop_k(optional): Number of results to return (default: 5)
search_collection(collection_name, query, top_k?, custom_tags?, filters?)
collection_name(required): Collection name or slug (usediscover_collectionsto find names)query(required): Natural language search querytop_k(optional): Number of results (default: 5)custom_tags(optional): Tags to scope retrievalfilters(optional): Metadata filters
Dynamic Collection Tools
For each collection you have access to, dedicated tools are generated using the collection's slug:
search_{slug}— semantic search with optional version, tags, and filtersget_topic_{slug}— retrieve information about a specific topiclist_versions_{slug}— list available documentation versions
For example, a collection called "Employee Handbook" gets:
search_employee_handbookget_topic_employee_handbooklist_versions_employee_handbook
Configure tool names and descriptions in the Ragora dashboard under Collection > MCP Settings.
Usage Examples
Search all collections:
"Search my docs for webhook event handling"
Search a specific collection:
"Look up the refund policy in the employee handbook"
Discover what's available:
"What workspaces do I have, and what's in them?"
How It Works
Your AI Tool Ragora Cloud
┌──────────────────┐ ┌──────────────────┐
│ Claude / Cursor │ HTTP POST │ MCP Server │
│ / VS Code / etc │──────────────>│ (Streamable HTTP)│
└──────────────────┘ /mcp └────────┬─────────┘
│
┌────────▼─────────┐
│ Gateway API │
│ (Auth + Search) │
└────────┬─────────┘
┌────┴────┐
▼ ▼
Postgres Qdrant
(metadata) (vectors)
- No local install — connect directly via HTTP
- Multi-tenant — your API key determines which collections you can access
- Dynamic tools — tools are generated based on your accessible collections
- Hybrid search — dense (E5-Large) + sparse (BM25) with reciprocal rank fusion
Security
- API keys authenticate all requests via
Authorization: Bearerheader - Each API key is scoped to the user's owned and purchased collections
- Cross-tenant data access is impossible — all queries are filtered by accessible collection IDs
- API keys should never be committed to version control
Troubleshooting
| Issue | Solution |
|---|---|
| "Tool not found" | Restart your AI app after config changes. Verify your API key starts with sk_live_ |
| "Unauthorized" | Check the Authorization: Bearer header format. Verify the key hasn't expired |
| No results | Ensure documents are uploaded and processing is complete. Try broader queries |
| Connection timeout | Check internet access. Verify server status at https://mcp.ragora.app/health |
Links
- Ragora — Sign up and manage your knowledge bases
- MCP Integration Guide — Full documentation
- API Keys — Create your API key
- Model Context Protocol — MCP specification
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。